Java Reference
In-Depth Information
G
,
H
Once we find an invalid foreign key, we buffer it and fail later, with the failure message
containing all violations. Although it would be simpler to fail right away, this approach
is more helpful, because it requires just one run to spot all invalid names.
The meaning of a valid foreign key is up to the project. In our case, we define that it
has to start with
FK_
and have two names (which should represent the tables involved
in the relationship) separated by an underscore (
_
).
I
This test case relies on a new infrastructure method,
analyzeSchema()
, which is
defined in
AbstractJpaTestCase
; listing 18.23 shows all changes required to imple-
ment it.
Listing 18.23
Changes on
AbstractJpaTestCase
to support
analyzeSchema()
[...]
public abstract class
AbstractJpaTestCase {
[...]
protected void
analyzeSchema( SqlHandler handler ) {
ConfigurationCreator cfgCreator =
new
ConfigurationCreator();
Configuration cfg = cfgCreator.createConfiguration();
SchemaExport export =
new
SchemaExport(cfg);
B
C
ByteArrayOutputStream outputStream =
new
ByteArrayOutputStream();
PrintStream oldOut = System.out;
PrintStream newOut =
new
PrintStream(outputStream);
System.setOut(newOut);
try
{
export.create(true, true);
String sql = outputStream.toString();
handler.handle(sql);
}
finally
{
System.setOut(oldOut);
newOut.close();
}
}
D
protected interface
SqlHandler {
void
handle( String sql );
}
E
private class
ConfigurationCreator
extends
JPAConfigurationTask {
@Override
protected
Configuration createConfiguration() {
return super
.createConfiguration();
}
}
analyzeSchema()
is based on the Template Design Pattern (see section 17.7.1). It
knows how to create a Java string containing the database schema but doesn't
know what to do with it, so it passes this string to a
SqlHandler
, which was passed
as a parameter.
SchemaExport
is the Hibernate class (part of the Hibernate Tools project) used to
export the schema. It requires a Hibernate
Properties
object, which unfortunately
B
C
